Condition monitoring is widely applied to determine a current and possible future status of mechanical machinery, including for example rotational systems such as turbines, generators, railway, and other industry processing equipment. Generally, data relating to temperature, rotation, vibration, etc, is analyzed for determining a change in operation, possibly initiating predictive maintenance of the rotating system for securing a long lifetime as well as reducing down time in case of equipment failure. There exist different techniques for performing condition monitoring of rotational systems, including frequency analysis of a vibration signal generated by a vibration sensor arranged in the vicinity of e.g. a rolling element bearing of the rotational system. In regards to e.g. rolling element bearings, the failure modes tend to be the degradation of those bearings, which typically exhibit an increase in characteristic frequencies associated with the bearing geometries and constructions that accordingly may be identified through frequency analysis.
An example of a monitoring system employing the above technique is disclosed in US2011/0020122 A1, describing a computer implemented method where condition information from a plurality of condition detection sensors coupled to a wind turbine are received by a centralized wind turbine controller. In US2011/0020122 A1, an anomaly detection algorithm is applied to identify maintenance activities for the wind turbine as a function of both the wind turbine condition information and the wind turbine controller information. Even though the monitoring technique disclosed in US2011/0020122 A1 provides a promising approach for reducing down time in a wind turbine, the use of a centralized controller puts high constraints on the reliability of communication within the monitoring system for securing essential real-time processing of the condition information. Such an approach will for example place limitations on the monitoring system in case future up-scaling of the system is desirable.
Accordingly, there is a need for improvements relating to distribution of condition information within a monitoring system, for allowing simplified up-scaling of the system in case of introduction of further condition detection sensors, as well as allowing for improvements in essential real-time processing of the condition information.